In a recently released issue of Pediatrics, co-first authors Drs. Helen Koechlin and Joe Kossowsky and colleagues present a “network meta-analysis” of non-pharmacologic treatments for pediatric migraine (10.1542/peds.2019-4107). Their objective was, “…To examine whether non-pharmacological treatments are more effective than waiting list (as a comparator) and whether there are differences between interventions regarding efficacy…” for children and adolescents with migraine. Twelve studies with a total of 576 participants (mean age 12.01 years, 61% female) were compared in this systematic review and meta-analysis. After reading this article, it was good news to learn how many non-pharmacologic treatments there are for pediatric migraine and to learn what a “network meta-analysis” is all about.
The authors identified studies of multiple different non-pharmacologic treatments for migraine. Treatments included relaxation, biofeedback, psychological treatments, and self- administered psychological treatments. Within each category there were several unique approaches: for example, the larger category of “biofeedback” included “biofeedback, biofeedback stress management, biofeedback relaxation education, and autogenic feedback (a hypnosis-based mental healing method).” The authors’ analyses included both the “lumping” approach, in which each larger category included several related clinical treatments, and the “splitting” approach, in which each unique treatment was compared to others. Studies of two potentially relevant treatments, music therapy and acupuncture, could not be compared using the network analysis approach.
So, what is network meta-analysis? “Network meta-analysis is a technique for comparing three or more interventions simultaneously in a single analysis by combining both direct and indirect evidence across a network of studies.”1 Usually an experimental intervention or treatment is contrasted directly with a comparator intervention, however, sometimes several interventions are compared within the same study, and one would like to understand which is the best of the alternatives, and further, compare this across multiple studies. That’s when network meta-analysis can be especially helpful, because it can both directly compare interventions within randomized controlled trials and indirectly compare across trials at the same time.2 More information is available in the cited references, and you might want to zero in on Figure 11.1.a1to see a diagram of how this information is displayed: “nodes” represent treatments, and the connecting lines represent (2-way or 3-way or even 4-way) comparisons between these treatments. The figures in the article by Koechlin, Kossowsky, and colleagues are incredibly helpful to understanding the study results.
I don’t want to spoil all of your reading fun, but the results are fascinating. When the authors “lumped” non-pharmacologic treatments, it was easy to see that each category of treatment was effective for pediatric migraine. Interestingly, when they “split” the treatments and compared each unique type, the statistically significant effects disappeared. Read more to understand why this may have happened, what it means, and where future research could take this field. My take-home for my patients, though, is that there are many options I have yet to explore beyond my usual first line approach of ibuprofen, sleep hygiene, self-care and trigger hunts. I hope you enjoy this article as much as I did.
References
- Chaimani A, Caldwell DM, Li T, Higgins JPT, Salanti G. Chapter 11: Undertaking network meta-analyses. Cochrane Training. https://training.cochrane.org/handbook/current/chapter-11# Accessed 3/10/21.
- Li, T., Puhan, M.A., Vedula, S.S. et al. Network meta-analysis-highly attractive but more methodological research is needed. BMC Med 9, 79 (2011). https://doi.org/10.1186/1741-7015-9-79
